LAUSR

Abstract Temporal records of river flow as essential data for water resource management plans are not available in many rivers around the world. In this study, monthly discharges of the Taskoh river in the north of Iran were simulated using dendrohydrology (the earlywood vessels and tree- rings chronologies). For this purpose, cross-dating, standardization, and time series analysis were performed on chronology records. The monthly river discharges during the growing seasons were simulated using an artificial neural network (ANN) by establishing a relationship between chronology records and the monthly river discharges. The result showed the high performance of the ANN model for simulating the monthly river flow using the mean vessels’ diameter and the tree-rings width in training (R-squared = 0.98, MSE = 0.009) and test (R-squared = 0.86, MSE = 1.2) stages. Moreover, the results showed that the vessels’ diameter is a better parameter (R coefficient = 0.87) than tree-rings width (R coefficient = 0.68) in river discharge simulation. Finally, the monthly river flow during the trees growing season was reconstructed using the tested ANN, mean vessels’ diameter, and tree-rings data for the past decades.